Puneet K

Robotic process automation (RPA) is a preconfigured software instance that uses business rules and predefined activity choreography to complete the autonomous execution of a combination of processes, activities, transactions, and tasks in one or more unrelated software systems to deliver a result or service with human exception management. . Features of RPA are
 
Works at the server level, independent of human intervention (unattended automation). Fully automates end-to-end processes. Scalable across multiple systems and departments. Ideal for structured, rule-based processes.  
 
Robotic desktop automation (RDA): computer application that makes available to a human operator a suite of predefined activity choreography to complete the execution of processes, activities, transactions, and tasks in one or more unrelated software systems to deliver a result or service in the course of human initiated or managed workflow. It can also be called as agent-assist automation, assistive automation, in-line automation. It Operates on an individual's desktop. Features of RDA are
 
Requires human intervention (attended automation). Best suited for tasks where decision-making and manual oversight are needed. Limited scalability  
Below are the different stages for designing an automated tool or a robot
 
Discover; Structured approach to review process landscape and identify best candidates for automation Design - Simple rapid engagement approach to understand the process and design and implement an automated solution Manage - Central deployment and management of automations with key governance controls and strong change management Run- Automations execute the business processes they are designed to perform while being monitored Engage -Communicate with business process teams to ensure the process is performing and exceptions are reported Measure - Reporting of utilization and impact of automation  
We have to make choice between RPA and RDA at the discovers stage itself. This is also illustrated by the diagram below
 

 
If the process requires frequent human interaction, RDA is preferable. If the process is rule-based and can be fully automated without human input, RPA is the better choice. If scalability is a priority, RPA is more effective. If process mapping accuracy is a concern then we should start with RDA to refine the workflow before transitioning to RPA. The issue of complexity of the data can still be handled by RDA through below means  
Modular Design: By dividing complex processes into smaller, reusable modules, RDA can address different parts of a workflow independently, making it easier to manage and update specific components without affecting the entire process.  Conditional Logic: RDA tools allow for "if-then" statements and decision points within the workflow, enabling robots to adapt to changing situations and make choices based on specific conditions within the data.  Data Extraction and Manipulation: RDA can extract relevant data from various sources, clean, transform, and manipulate it as needed within the workflow, handling complex data structures and calculations.  Application Integration: By integrating with diverse applications and systems, RDA can automate cross-platform processes, pulling data from one system to perform actions in another, even if they have different interfaces.  Looping and Iteration: RDA can repeat certain steps within a workflow based on defined criteria, making it suitable for handling repetitive tasks with variable data input.  Error Handling: Advanced error handling mechanisms allow robots to identify and recover from potential issues during execution, automatically retrying steps or notifying users when necessary. 

A moonshot idea is an ambitious, groundbreaking concept aimed at achieving something that seems nearly impossible but could revolutionize an industry, society, or technology if successful. Moonshot ideas typically involve high risk, out of the box thinking, and possible significant investment in research and development.
 
 
For an idea to be called as Moonshot Idea it should have the below statistics:
 
1. Radical Innovation – It shouldn't be a simple fix rather it should be revolutionary, unimaginable, or undoable. It should be transformative in nature and not BAU
 
2. Based on principle of high risk * High reward- While risk will be eminent in terms of investment of time and money, reward won't be certain however if it is successful then rewards to be multiple x of investment made.
 
3. Should involve cutting-edge technology or scientific breakthroughs.
 
4. Global Impact – If successful, it can change the world / industry
 
Being in finance, I often encounter the challenges of external payments being prone to fraud and not safe. In the new digitized world, I would like to explore the idea of AI in smart payment. So far use of LLM in payments is rare and yet to have confidence from finance community.
 
Various checks which I would potentially have to do for moonshot idea are :
 
Technological Feasibility Is LLM matured enough to handle topics are payments whereby robots also struggle to keep the fraud risk down despite the pre-defined criteria Any organization who has experimented this in past and what were the outcomes and reasons for failure No human touch environment possibility thereby eliminating risk of internal fraud/ Fully automated AI enabled Human less Data Management Team. AI Powered Tools / Metrices Current accuracy % and error rate – May be even factor analysis could help in here Real Time payment incoming pairing alert from recipient server which if not provided would lead to stoppage of payment, ( Host to host) Possibility of block chain based decentralized fraud detection network across different regulatory environments Regulatory & Compliance considerations – Payment laws. Network limitations/ VPN regulations and country specific restrictions like China, Brazil etc. Cost of Prototype ( MVP) and final product needs to be justified. ( Business Value Analysis skills learnt during MBB Course will be truly tested) Simulation environment and test cases from one of the BUs / regions/ geographies Scaling up strategy- Partner with existing player or produce inhouse Exclusive use or sale in the market.  
As we can see there could be number of considerations which are to be made before we embark on journey of a moonshot idea.

Factor Analysis is often used to the reduce the challenges of complex data. It identifies the variables which are not apparent ( latent or underlying) which explains the cause of variances in data. It doesn't identifies number of individual variables rather it groups variables into a s a smaller set of factors and draws patterns which makes the relationship more understandable & obvious. This calculated reduction or selective approach helps in identifying core drivers of the varability and eliminates redundant information.
 
Factor analysis work in the following manner :
 
 
1. Examines and Identify Correlations among multiple observed variables.
 
 
2. Extract underlying Factors by grouping highly correlated variables.
 
 
3. Interpret Factors basis variables it relies highly on.
 
 
4. Reduce Dimensions of variables while retaining the most significant information.
 
Practical Example in a Lean Six Sigma Project
 
Scenario: Reducing Customer Complaints in a customer grievance Center
 
A Lean Six Sigma team is analyzing customer complaints in a call center to improve service quality. They collect survey responses on 10 different service attributes:
 
1. Response Time
 
 
2. Agent's behaviour
 
 
3. Call Resolution
 
 
4. Clarity of Information
 
 
5. Subject Matter Knowledge of the Agent
 
 
6. Call Hold Time
 
 
7. Service Process Efficiency
 
 
8. Ease of Reaching the agent
 
 
9. Follow-Up Effectiveness
 
 
10. Call Escalation
 
 
 
Applying Factor Analysis
 
The analysis reveals that these 10 attributes can be grouped into three main factors:
 
Factor 1: Service Efficiency (Response Time, Hold Time, Ease of Reaching agent)
 
Factor 2: Agent Performance (Agent behavior, Knowledge, Call Resolution)
 
Factor 3: Process Effectiveness (Follow-Up Effectiveness, Call Escalation, Clarity of Information)
 
Lean Six Sigma Benefits
 
Simplifies analysis: Instead of analyzing all the 10 variables separately, the team focuses on three key drivers.
 
Prioritization of improvements: If Factor 1 (Service Efficiency) has the highest impact on customer complaints then that becomes a focus area, and the efforts will be done to improve it.
 
Reduces redundant efforts: The team avoids fixing individual variables separately and instead works on holistic improvements in identified factors.
 
By using factor analysis, the Lean Six Sigma team can streamline the problem-solving process, leading to more effective decision-making and efficient resource allocation